摘要
为了提高视频水印的鲁棒性,提出一种基于运动目标检测和纹理复杂度相结合的视频水印算法。通过在线聚类的背景减法和背景差分法提取出运动目标来生成水印,并用混沌加密的方法对生成水印进行加密。在对宿主帧进行分块的基础上将加密后的水印嵌入到纹理复杂度较大的分块中。实验表明,该方法在保证视频图像质量的前提下对于常见的攻击有良好的鲁棒性。
In order to improve the robustness of the video watermarking, a video watermarking algorithm combining the moving target detection and the texture complexity was proposed. The watermarking was generated by using online clustering of the background subtraction and the background difference to extract the moving target. The generated watermarking was then en- crypted by the chaotic encrypted method. Based on the block of the host frame, the encrypted watermarking was embed into a block with larger texture complexity. The experiment results showed that the good robustness for the common attacks was obtained under the premise of ensu- ring the quality of the video images.
出处
《西安邮电学院学报》
2013年第1期35-39,共5页
Journal of Xi'an Institute of Posts and Telecommunications
基金
西安市科技计划创新支撑类基金资助项目(CX1129(n))
关键词
运动目标检测
纹理复杂度
混沌加密
自嵌入
moving target detection, texture complexity, chaotic encryption, self-embedding